Relapse remains the major obstacle to cure following hematopoietic cell transplantation (HCT). Despite preparative regimens employing pharmacokinetic targeting of chemotherapy, radioactive antibodies, or adjunctive immunotherapy, relapse remains the leading cause of treatment failure following HCT. Indeed, a recent NCI sponsored conference addressed the problem of relapse post-HCT, and among the conclusions were that studies of pre- and post-transplant samples needed to be performed to understand the biology of relapse (1). This proposal addresses this issue. Thus in Aim 1 we will determine the genetic predictors of outcome following hematopoietic cell transplantation (HCT) for acute myeloid leukemia (AML). We hypothesize that there are genetic pathways that predict treatment response to HCT independent of disease stage (defined by blast count and cytogenetics). Thus, we will use mRNA and miRNA expression analysis to identify genes and pathways that distinguish patients who fail, i.e. relapse after HCT and patients who are disease-free following HCT. The results of this aim will allow us to better risk stratify patients to diffeent treatment approaches, as well as give us insight into the biological mechanisms that drive response following transplantation. In Aim 2 we will define the genetic changes in AML during minimal residual disease (MRD) and relapse. We have developed methods that can perform gene expression on small numbers of cells captured with flow cytometry. Thus, we will follow the gene expression and of the signature discovered in Specific Aim 1, as well mutational genotype, at diagnosis, MRD, and relapse (if this occurs). This will allow us to refine MRD detection to understand not only how much residual disease remains post-therapy, but define its molecular characteristics, which likely will be important to both predicting relapse, as well as selective pre- emptive therapy. Lastly, in Aim 3 we will define the role of clonal selection in AML post-HCT relapse. This aim will compare diagnostic and relapse samples by single cell genotyping and gene expression to understand how the relapsed sample compares to that of the pre-transplant disease. An understanding of the context and extent of clonal selection in relapse may prove important in tailoring treatment strategies to minimize selection and the escape of resistant clones.